Spark decimal type precision loss

This is a current open issue, see SPARK-27089. The suggested work around is to adjust the setting below. I validated that the SQL statement works as expected with this setting set to false.

spark.sql.decimalOperations.allowPrecisionLoss=false